Machines Can’t be Held Accountable

In recent years, cars have been developing more and more technology that claims to make them “safer,” such as lane assist technology, automatic braking, and perhaps the most impressive, full self driving. All of these technologies, and many more, claim to make cars safer, while actually just making drivers pay less attention to the road. Self driving is the worst of all of these, as it allows drivers to ignore the road altogether, letting an unreliable system take control of their car, and by proxy, the lives of everyone in and around it.

There are two main ways that cars can sense the world around them to enable self driving. One involves cameras, often using AI to process the images, which is the method newer Teslas use. This can be good because it allows the car to tell a difference between different objects, for example a truck trailer and a wall. It does, however, require a lot of processing power and doesn’t work as well in severe weather such as fog or rain. The other option uses depth sensing technologies, such as LIDAR or ultrasonic sensors. Older Teslas used ultrasonic sensors, and most newer cars use LIDAR. Both can detect objects, even through fog, rain, and other weather. The downside is that they cannot tell the difference between types of objects, however this isn’t really necessary for the most part. The car can see an object somewhere and know not to drive into it, regardless of what it is.

One big issue with self driving cars is that, like many other “safety” features in modern cars, they take the driver’s attention away from the road. Even if you still pay attention to the road while the car is driving (which you should) you’re still not actually controlling the vehicle, and aren’t as involved with the driving. This leads drivers to be less likely to be able to react quickly in the event of the autopilot disabling or malfunctioning, and also raises the question of whether the driver or the car is responsible for any damages caused by the autopilot.

Another concern, and probably the biggest, is that autopilot systems can never be perfect, no matter what type of sensors it uses. There will always be moments where an autopilot system misinterprets something, from something as small as not slowing down for a speed bump to something as dangerous as driving into a person, car, or building. This is something that will likely get better over time, but will never be perfect. Cars and roads were built for humans to be able to navigate, and the only way self driving cars will be able to confidently direct themselves would require completely re-imagining the world’s traffic system.

In conclusion, self-driving cars are an impressive showcase of what computers and cars can do, but like any other driver assistance system, all it really does is take drivers’ attention away from the roads and make driving more dangerous.

“Toyota self-driving car – Consumer Electronics Show – CES 2013 – Las Vegas” by David Berkowitz is licensed under CC BY 2.0.